Mario Hernandez

1 · 2025

Read Mario Hernandez by 1 (2025) in the Tower of Babel, a free online library of public-domain works preserved in clean, searchable text. No sign-up. The full text is served as a plain-text file at a predictable URL, where text is present: https://antarctic-labs.com/tower-of-babel/arxiv-2509-14523.txt.

Introduction

Mario Hernandez and four co-authors build a radio test bed in 2025 for sharing LiDAR point clouds between vehicles. Cars that pool sensor views can see around corners and occlusions together. Testing vehicle to vehicle links needs hardware between simulation and road deployment.

Their platform uses software defined radios running the vehicular WiFi standard, built on low cost Pluto devices. Any computer running the container, Matlab and a radio over USB can act as a node. LiDAR clouds are shared and fused into a collective perception view in the demonstration. The team also models decentralised storage options and studies channel quality. A cheap modular test bed lets more labs debug cooperative sensing before road trials.

Catalog record

CREATOR1
YEAR2025
COLLECTIONPAPERS
FORMATTXT
FILE SIZE20 KB
VERSION1.0
LICENSEhttp://creativecommons.org/licenses/by/4.0/
RIGHTSCC BY
SOURCEarXiv

Source: https://arxiv.org/abs/2509-14523

Browse the full library · About the Tower of Babel

Full text

The opening of the work, from the library's hosted text:

Mario Hernandez 1 Abstract We present a Software Defined Radio (SDR)–based IEEE 802.11p testbed for distributed Vehicle-to-Vehicle (V2V) communication 1 1 1 Code: here . . The platform bridges the gap between network simulation and deployment by providing a modular codebase configured for cost-effective ADALM-Pluto SDRs. Any device capable of running a Docker with ROS, executing Matlab and interface with a Pluto via USB can act as a communication node. To demonstrate collaborative sensing, we share LiDAR point clouds between nodes and fuse them into a collective perception environment. We evaluated a theoretical model for leveraging decentralized storage systems (IPFS and Filecoin), analyzing constraints such as node storage convergence, latency, and scalability. In addition, we provide a channel quality study. A Software-Defined Radio Testbed for Distributed LiDAR Point Cloud Sharing with IEEE 802.11p in V2V Networks 1. Introduction Vehicle-to-Vehicle communication is a key enabler of intelligent transportation systems and autonomous driving. By communicating environmental information, vehicles can improve situational awareness, enhance safety and coordinate maneuvers. Standards such as IEEE 802.11p provide the physical (PHY) and medium access control (MAC) layers needed for Dedicated Short-Range Communication (DSRC), but significant challenges remain in evaluating real-world performance of perception sharing across heterogeneous nodes. Recent advancements in blockchain-based storage such as InterPlanetary File System (IPFS) and Filecoin have introduced decentralized, incentivized, tamper-resistant methods for sharing large datasets across distributed networks. This addresses a critical safety issue in these multi-node infrastructure networks, where data could be easily read and tampered with by a knowledgeable party. For V2V communication systems, these decentralized peer-to-peer storage methods offer the potential to offload and archive high-bandwidth perception data, including LiDAR point clouds, while ensuring verifiability, resilience against single points of failure, and availability to nodes without direct internet connectivity. Our main contributions are as follow: • We developed a Dockerized, lightweight ROS base node that is fully extensible, enabling heterogeneous devices to join the network with minimal setup. • We integrate MATLAB-based SDR communication with ROS, enabling real-time fusion of LiDAR point clouds through ROS topics. • We give a theoretic analysis of data propagation throughout the network and per-node storage constraints. • We provide a channel quality study. 2. Related Work Cooperative perception has advanced through simulation frameworks such as OPV2V [ 1 ] and infrastructure-based LTE/C-V2X studies [ 2 ] . These works highlight the benefits of multi-view fusion and provide scalable environments for testing algorithms, though their simulation based nature limits realism. Our testbed complements this line of research by enabling real-time SDR-based LiDAR sharing across ROS-managed nodes, providing a physical platform for validating collaborative perception. SDRs have been widely used to prototype systems such as IEEE 802.15.4 [ 3 ] , LTE [ 4 ] , and WLAN [ 5 ] , ranging from open-source to closed-source. These efforts typically address PHY/MAC layer communication but don’t address containerization for general node types and backend data processing (e.g. point-cloud fusion). We complement this body of research by adapting 802.11a/g reference designs [ 5 ] to 802.11p and pairing them with Dockerized ROS nodes for V2V perception sharing. Additionally, we leverage Matlab’s opensource software and plug-and-play capabilities. Decentralized vehicular networking has been explored in areas such as blockchain security [ 6 , 7 ] and peer-to-peer file sharing [ 8 ] . These studies emphasize scalability, trust and highlight the potential for secure vehicle-to-vehicle communication leveraging blockchain technology. 3. System Design and Implementation Figure 1 : System overview of the IEEE 802.11p SDR V2V communication system. A single information pass shows how a node transmits and receives LiDAR point-cloud data and fuses it via ROS topics. Our SDR-based V2V testbed is designed to be highly configurable and modular, supporting multiple modes of operation to accommodate different experimental setups and objectives. The system has four main modes: • Transmit (TX) – Continuously sends generated LiDAR point cloud data over the air to other nodes. • Receive (RX) – Listens for incoming transmissions from neighboring nodes and reconstructs the LiDAR data. • Self-Transmit – Allows a single node to transmit and immediately receive its own signals for calibration or testing purposes. • Discovery – Implements ALOHA-based medium access algorithm [ 9 ] , constantly listens but periodically broadcasts a presence signal at random intervals so that nodes can detect and identify each other in the network without message collisions. These modes allow researchers to simulate real-world V2V scenarios, including one-to-one, one-to-many, and many-to-many interactions, without modifying the underlying hardware. The system is modular: both Self-Transmit and Discovery function as loops using the TX and RX modes as building blocks. The system integrates Matlab for SDR signal control and ROS containers for backend LiDAR data fusion. This architecture supports experimentation with distributed networks of vehicles, drones, or any node capable of running ROS in a Docker container and interfacing with SDR hardware. 3.1 SDR Hardware Choice We use cost-effective ADALM-Pluto SDRs as the hardware backbone. SDRs provide full control over PHY and MAC layers, making them ideal for research and experimentation. These allow is to configure DSRC-compliant transmissions by adjusting parameters such as channel frequency, gain, and packet scheduling. This flexibility allows us to configure multiple network protocols through software (in our case Matlab) instead of having specialized automotive radios. 3.2 IEEE 802.11p The choice of IEEE 802.11p [ 10 ] or DSRC (Dedicated Short-Range Communication) is natural since it is the established standard for low-latency, short-range vehicular communications and most radios in "the wild" today aren’t capable of communicating in more sophisticated CV2X or 5G formats. DSRC is specifically designed to support high-mobility environments and offers predictable latency and reliable packet delivery. IEEE 802.11p was introduced as an amendment to IEEE 802.11-2012 to enable communication outside the context of a basic service set (OCB) . This eliminates time-consuming authentication and association, which are impractical in highly mobile vehicular environments. Instead, frames are transmitted with a wildcard BSSID, allowing true ad hoc V2V operation. PHY Layer. The 802.11p PHY uses orthogonal frequency division multiplexing (OFDM) in the 5.9 GHz band with 10 MHz wide channels, halving the 20 MHz channelization of WLAN to improve Doppler tolerance. Each channel supports eight data rates from 3 Mbit/s to 27 Mbit/s, using BPSK, QPSK, 16-QAM, and 64-QAM with coding rates from 1/2 to 3/4 (see Table B.1 in [ 10 ] ). Our implementation uses MCS 4, 16-QAM modulation with OFDM and code rate 3/4, providing 18 Mbit/s while maintaining resilience to multipath fading. OFDM enables robust multipath performance, critical for driver environments, while the choice of 16-QAM balances throughput and error resilience for LiDAR point-cloud transmissions. MAC Layer. The 802.11p MAC is derived from IEEE 802.11e Enhanced Distributed Channel Access (EDCA), itself an extension of CSMA/CA with QoS support (refer to B.4.1 in [ 10 ] ). In this scheme, a node performs carrier sensing and defers transmission by a random back-off drawn from a contention window. The window expands exponentially upon collisions and resets after successful transmissions. For broadcast V2V messages, acknowledgements (ACKs) are suppressed, so the backoff is only applied once per packet and the contention window remains at CW_{min} . This ensures low latency but provides no retransmission, a trade-off consistent with safety-critical broadcast beacons. Implementation note. While our

Read the complete text